Original ContributionPulse Wave Imaging in Carotid Artery Stenosis Human Patients in Vivo
Introduction
Atherosclerosis is a chronic vascular disease characterized by compositional changes in the arterial walls that lead to the buildup of plaque, which consists of lipids, cholesterol, calcium and other substances found in the blood (Libby et al., 2011, Mahmoud et al., 2013). Carotid stenosis is a narrowing of the lumen in the carotid artery usually caused by atherosclerosis, occluding blood flow to the brain. A stroke may occur if the plaque ruptures and forms a blood clot (i.e., cerebral thrombosis), which may detach and become lodged upstream in the smaller vessels of the brain (i.e., cerebral embolism). More than 15 million people suffer strokes each year worldwide, resulting in ∼5 million deaths (Mughal et al. 2011). In the United States, stroke affects approximately 795,000 people each year and is the third leading cause of death (>140,000 annually). An estimated 15% to 20% of all ischemic strokes are attributed to carotid atherosclerosis (Taussky et al. 2011).
In patients exhibiting severe carotid blockage, a surgical intervention such as a carotid endarterectomy (CEA) or carotid stenting may be performed to reduce the risk of stroke. Current clinical practice for selecting patients for a CEA is heavily based on symptomatology and the degree of stenosis (Chan et al. 2014). However, the majority of ischemic stroke cases occur because emboli originating from a carotid plaque occlude an artery supplying the brain, not because of the luminal narrowing itself (Moller et al. 2012). Thus, a significant diameter reduction may not always correlate with a high risk of stroke. In fact, histopathological studies have found that cerebrovascular events can also occur in patients with carotid plaques causing low-grade stenosis (<30%) and with no other identifiable cause for their stroke (Lovett et al., 2004a, Lovett et al., 2004bb; Wasserman et al. 2005). As a result, nearly 80% of the ischemic strokes attributed to carotid atherosclerosis occur in asymptomatic patients without a history of stroke or transient ischemic attacks (i.e., “mini-strokes”) (Taussky et al. 2011). For symptomatic patients with <70% stenosis and for asymptomatic patients, the degree of stenosis alone may not be a reliable measure of stroke risk (Xu et al. 2014). Identification of patients with high-risk asymptomatic carotid plaques remains an elusive but essential step in stroke prevention.
Several recent investigations have found that the plaque structure and composition may represent a more direct biomarker for the development of cerebrovascular ischemic events rather than the degree of luminal stenosis (Naghavi et al., 2003, Saba et al., 2014). Thus, the early detection of carotid atherosclerotic disease and reliable identification of plaque features associated with an increased risk of rupture are crucial for stroke prevention. Atherosclerotic plaques can be broadly categorized into vulnerable and stable (Finn et al. 2010). Vulnerable (i.e., unstable) plaques are often characterized by a thin fibrous cap covering a large necrotic core containing macrophages and interstitial collagen (van den Oord et al. 2014), leading to an increased risk of rupture. Stable plaques tend to be asymptomatic (Ross 1993) and are characterized by an intact and thick fibrous cap consisting of smooth muscle cells in an extracellular matrix rich in type I and III collagen (Finn et al. 2010). Calcification is common in late-stage atherosclerotic plaques and increases with age (Bentzon et al. 2014). The combination of calcium deposition and the collagen-rich matrix increases the stiffness of the plaque and contributes to its stability.
Non-invasive imaging methods to assess the correlation between plaque properties and risk of stroke have been developed primarily using magnetic resonance imaging (MRI), computed tomography (CT) and ultrasound techniques (Chan et al., 2014, Underhill et al., 2010, Xu et al., 2014). For example, calcified plaques are depicted as high-intensity regions on CT angiograms, while ultrasound methods include characterization of plaques based on their acoustic properties (Brewin et al. 2014) and echogenicity (Moller et al. 2012). Studies comparing plaque histology with ultrasonography have suggested that echolucent plaques tend to be higher in lipid content and echogenic plaques contain more calcified and/or fibrous tissue (Gonçalves et al. 2004).
The majority of research on imaging-based methods to assess plaque properties focuses on identifying features from a single image rather than investigating the changes in vascular dynamics associated with different types of plaque. Pulse wave imaging (PWI) is a previously developed ultrasound elasticity imaging-based technique for the spatiotemporal mapping of pulse wave-induced arterial wall motion (Apostolakis et al., 2017a, Fujikura et al., 2007, Li et al., 2013, Luo et al., 2009, Nandlall and Konofagou, 2016, Vappou et al., 2010), facilitating the measurement of local pulse wave velocity (PWV). It should be noted that although PWV is known to correlate with arterial stiffness, direct indices of arterial stiffness (i.e., modulus or compliance) are also affected by geometric parameters such as diameter and thickness (Holzapfel, 2006, Khamdaeng et al., 2012, Westenberg et al., 2012), which can be measured from the PWI images. The modified Moens–Korteweg equation (Korteweg 1878; Moens 1878; Fung 1997) has traditionally been used to relate PWV to the incremental elastic modulus; however, assumptions such as an infinitely long, straight, isolated and cylindrical vessel with elastic, isotropic and homogenous walls, containing a homogenous, incompressible and non-viscous fluid are compromised by carotid stenosis. Also, the wave speed (i.e., PWV) in the Moens–Korteweg equation represents the speed of the wave propagation with respect to the fluid, which may not be represented by the wave propagation in the wall under turbulent flow conditions. The Bramwell–Hill model (Westenberg et al. 2012) proposed a series of substitutions relevant to the observable hemodynamic measures and was used to derive the compliance of the normal controls and each plaque region:where D is the diameter (measured at each scan line position), and ρ is the fluid density of blood (∼1060 kg/m3).
Early PWI studies in CaCl2 and AngII-induced abdominal aortic aneurysm (AAA) mouse models (Fujikura et al., 2007, Luo et al., 2009) revealed that the uniformity of the pulse wave propagation may serve as a valuable biomarker to differentiate between normal and diseased arteries. The feasibility of PWI in the carotid artery in vivo has been reported in young, healthy patients (Luo et al. 2012). PWV measurements at a segment of the left common carotid artery (CCA) away from the bifurcation in eight male volunteers ranged from 4.0 to 5.2 m/s. More recently, PWI was implemented with plane wave acquisitions that are subsequently coherently combined using coherent compounding, thus improving the accuracy and reliability of the PWV measurements at high temporal and spatial resolution (Apostolakis et al. 2017a). With this technique, good reproducibility among six healthy patients was found in the estimated carotid artery PWVs over the course of 1–3 d (first acquisition: 3.97 ± 1.21 m/s, second acquisition: 4.08 ± 1.15 m/s). Four-dimensional PWI utilizing a 2-D array transducer with plane wave acquisitions at 2000 Hz was also introduced recently and tested in phantoms and in vivo in healthy patients (Apostolakis et al. 2017b). Furthermore, the arterial wall displacements themselves have been used to derive the stress–strain relationship in vivo, illustrating the complex mechanical interaction of the different wall constituents (Khamdaeng et al. 2012).
Strain imaging, another ultrasound-based elastographic method that involves computing the spatial gradient of the displacement map to estimate strains, provides a new level of information regarding tissue elastic properties that is an active field of research in ultrasound imaging (Zaleska-Dorobisz et al. 2014). In the context of ultrasound elasticity imaging, strain represents the deformation of soft tissue in response to an external force (Ophir et al. 1999), such as the intra-luminal pressure of a pulsating artery acting on the walls in the radial direction. When this force deforms a medium such as a heterogeneous plaque, stiffer (softer) regions in the medium are expected to experience a lower (higher) level of strain. One- and two-dimensional ultrasound strain imaging has been attempted in carotid plaques (Hansen et al., 2016, Naim et al., 2013, Poree et al., 2015, Roy Cardinal et al., 2017), illustrating both the feasibility and limitations of using peak systolic cumulative strains for plaque characterization. One of the key considerations for robust arterial wall strain estimation is the size of the strain kernel, which must be small enough to operate within the structure of interest but large enough to ensure an adequate strain signal-to-noise ratio (Bunting et al. 2014). This is why strain estimation in a thin structure such as the normal carotid artery wall typically can yield noisy estimates. However, the increased thickness in atherosclerotic regions allows for the use of larger strain kernels to achieve higher signal-to-noise ratios.
Strain estimation using ultrasound has exhibited great promise for clinical integration because of its non-invasiveness, low cost and ease of use. Two-dimensional (i.e., both axial and lateral) strain estimation techniques have been developed by our group for cardiac imaging applications such as myocardial elastography (Konofagou et al., 2002, Lee et al., 2007, Lee and Konofagou, 2008) and mapping of the electromechanical wave (Provost et al. 2011).
Thus, combining the results of PWI and strain imaging may lead to new insights into how arterial function is affected by plaques of various composition, size and location. The PWV in a stenotic carotid segment as well as the pulse wave-induced displacements and strains, may be useful in detecting and characterizing plaque regions. In the work described here, the PWI and strain imaging methods were applied in patients with carotid atherosclerosis to investigate the effects of stenosis on local arterial mechanics.
Section snippets
Data acquisition
All imaging procedures were approved by the institutional review board of Columbia University Medical Center. Eleven patients (N = 13 plaques; 9 males, 2 females; mean age: 76.00 ± 8.51 y), diagnosed with moderate (>50% occlusion) to high-grade (>80% occlusion) carotid stenosis by a clinical expert, provided informed consent to participate in the present study. Radiofrequency signals were acquired using conventional ultrasound (SonixTouch, Analogic Corp., Peabody, MA, USA) and/or plane wave
Results
Figures 2 and 3 illustrate the displacement and single-depth PWV mapping results in two patients (72-y-old female and 76-y-old male, respectively), obtained using the conventional imaging sequence. In both cases, decreased cumulative displacements (Figs. 2b and 3b) were observed in the plaque of interest (red contour). Also in both cases, the piecewise PWV map (Figs. 2c, d and 3c, d) reveals a region where the PWV transitions from positive (i.e., proximal to distal) to negative (i.e., distal to
Discussion
Non-invasive methods that can reveal new information regarding the composition and stability of carotid plaques may play a key role in plaque characterization and stroke prevention. In this study, the feasibility of PWI was evaluated in patients with moderate to severe carotid stenosis. To increase the resolution of the wave propagation analysis for the detection and characterization of regional lesions, piecewise estimation of the PWV was used. The high spatial and temporal resolution
Conclusions
The results from this pilot clinical study indicate the potential of PWI to differentiate between plaques of varying stiffness, location and composition based on the cumulative displacements, cumulative strains, PWV and compliance. Characteristics such as pulse wave convergence, decreased strain and alternating positive and negative PWVs within the plaque were observed and may serve as valuable information to compensate for the limitations of methods currently used for the assessment of stroke
Acknowledgments
This work was supported by a grant from the National Institutes of Health (NIH R01-HL098830). I.Z.A. was also supported by the Gerondelis foundation.
References (77)
- et al.
Carotid atherosclerotic plaque characterisation by measurement of ultrasound sound speed in vitro at high frequency, 20 MHz
Ultrasonics
(2014) - et al.
Imaging of the vulnerable carotid plaque: Biological targeting of inflammation in atherosclerosis using iron oxide particles and MRI
Eur J Vasc Endovasc Surg
(2014) - et al.
Distension capacity of the carotid artery and ambulatory blood pressure monitoring: Effects of age and hypertension
Am J Hypertens
(1995) Determination of material models for arterial walls from uniaxial extension tests and histological structure
J Theor Biol
(2006)The carotid bifurcation plaque—A model for the study of atherosclerosis
J Vasc Surg
(1986)- et al.
Arterial stiffness identification of the human carotid artery using the stress–strain relationship in vivo
Ultrasonics
(2012) - et al.
Myocardial elastography—A feasibility study in vivo
Ultrasound Med Biol
(2002) - et al.
Angle-independent and multi-dimensional myocardial elastography—From theory to clinical validation
Ultrasonics
(2008) - et al.
Direct numerical simulation of transitional flow in a stenosed carotid bifurcation
J Biomech
(2008) - et al.
Performance assessment of pulse wave imaging using conventional ultrasound in canine aortas ex vivo and normal human arteries in vivo
Artery Res
(2015)